[FS#1259] PPPoE disconnects under high upstream load

[LEDE Bugs]

A new Flyspray task has been opened.  Details are below. 

User who did this - llucz (llucz) 

Attached to Project - LEDE Project
Summary - PPPoE disconnects under high upstream load
Task Type - Bug Report
Category - Base system
Status - Unconfirmed
Assigned To - 
Operating System - All
Severity - Medium
Priority - Very Low
Reported Version - Trunk
Due in Version - Undecided
Due Date - Undecided
Details - Hello,

I'm using an OpenWrt (Barrier Breaker r39582 ) TP-Link router to connect to the Internet through a DSL modem, via PPPoE.

Everything is fine, except when under high **upstream** load (i.e., when I saturate my upstream bandwidth, and only under these conditions). In these conditions, the OpenWrt router is unable to receive modem replies to PPPoE LCP echo-requests, and terminates the PPP connection. This behavior is experienced //only// under heavy upstream traffic, and is pretty deterministic.

This defect is also witnessed by the following log excerpt:

Sat Jan  6 17:47:13 2018 daemon.info dnsmasq[13213]: using nameserver 8.8.4.4#53
Sat Jan  6 17:47:13 2018 daemon.info dnsmasq[13213]: using nameserver 8.8.8.8#53
Sat Jan  6 17:47:13 2018 daemon.info dnsmasq[13213]: using local addresses only for domain lan
Sat Jan  6 17:50:01 2018 daemon.info pppd[12884]: No response to 6 echo-requests
Sat Jan  6 17:50:01 2018 daemon.notice pppd[12884]: Serial link appears to be disconnected.
Sat Jan  6 17:50:01 2018 daemon.info pppd[12884]: Connect time 2.9 minutes.
Sat Jan  6 17:50:01 2018 daemon.info pppd[12884]: Sent 53819998 bytes, received 1176989 bytes


I've attempted to play with PPPD lcp-echo-interval and lcp-echo-failure parameters, but I'm unable to solve the problem through these settings.

I guess the correct solution should be to "prioritize" (QoS?) PPP LCP echos, but maybe there is a simpler way to do it.

Thanks!
 

More information can be found at the following URL:
https://bugs.lede-project.org/index.php?do=details&task_id=1259